wizzwizz4

In the US, where race is entirely defined by skin pigmentation and there's heavy segregation ("redlining") of living quarters, sure: databases aren't necessary. But in places where race is defined by heritage or culture, things which can be disguised by the targets of a regime, these databases are a vital part of any serious ethnic cleansing operation. I don't remember the statistic in IBM and the Holocaust that compares Jew survival rates in occupied countries with versus without digitised censuses, but the difference was stark.

Italian fascism originated in Italy, but if you broaden the definition of "fascism" to include Nazism (as Adolf Hitler did), then there's an argument that it originated in the 19th century, in many parts of the world, including the United States. Among others, Adolf Hitler took inspiration from Madison Grant, Henry Ford, and the Jim Crow laws; and the Nazi regime took inspiration from Harry Laughlin, the KKK, and American eugenics. (WWII might be a bigger contributor to eugenics going out of fashion in the US than its lack of scientific merit was.)

Well, it's certainly impenetrable.

The user agent can refuse to store cookies, but it can't do much against supercookies (cookie-like features not knowingly implemented by the user agent programmers) or fingerprinting: you need something like legislation to curb practices like that.

Without further violating privacy. I would say that the data these companies are already collecting on people (including children, who many believe should be afforded extra protections from such companies) is a major, systematic privacy violation.

Mathematics is to hallucinatory mushroom people as the flap of a hummingbird's wings is to the scratches on a lens.

Yes. All the time the rest of the organisation spends on mandatory ethical training, the ethics board spends on martial arts.

I have a camera. Suppose I rub course-grained sandpaper all over its lens. Now the camera sees scratches, instead of the outside world. But those scratches are real! We can see their physical counterparts if we study the camera. Clearly, the camera has been much improved.

I may not be personally familiar with the latest and greatest moving target in agentic AI, but I am familiar with AI-generated "slop code". What I see is consistently defective, to the point where I would not trust an AI-generated data pipeline to produce remotely accurate results, even though messing up a basic data pipeline that much is a difficult task for humans. (Then there's the tendency for an LLM post-processing the tool output to smooth over such flaws, massaging or fabricating conspicuously missing or corrupt data to hide what would normally be extremely obvious warning signs.) If it is, as the proponents claim, even possible to entice these systems to produce decent programs (other than by direct plagiarism), it must require skill way beyond that of the average prompter with 18 months' experience to explain my observations.

When I say "write an ordinary computer program". I mean just writing the program, in a programming language. Your comment expresses disagreement with me, so should have started with "No", not "Yes".

> Try some context engineering. Try customizing your harness. Try having the harness improve itself. These are AI 101 lessons

If AI is really that complicated, it sounds like it would be easier to write an ordinary computer program to aggregate job boards.

Wishing for something doesn't make it happen. So we should consider the effect that this wishing has on you: does wishing for vengeance make you a worse, or better, person? I'd try to avoid it, myself.

If you're in a position where you're actually making the decision between letting these harmful practices continue, or putting a stop to them, that's different. But in this society, such people are more likely to be journalists and justices than vigilantes, so violence is still inappropriate. I think the scenario you described, of those incited to violence turning on the inciters with violence, is rather unlikely in reality, and it wouldn't necessarily be a good thing.

> HN pretends there's nothing to discuss here.

Have you read the discussions on vaccine stories? See, for example, https://news.ycombinator.com/item?id=26629234 – chosen arbitrarily using the search box at the bottom of the page. I've not read anyone pretending this.

> shut up

People are being dismissive of you, but nobody has told you to shut up in this thread. (The closest thing is flagging the comments which started with broad-spectrum insults, but this only occurred after you wrote this.) Despite its tone, I'm going to have to second GP's assessment that you're responding to imagined slights.

Does your television not take 10 seconds to start, 2 seconds to respond to button inputs, request internet access, try to sell you things, and exfiltrate your personal data? If not: where are you getting your televisions?

This kind of market shopping is common in grant-funded domains. There's a project that you know is a good, but nobody wants to pay for it… but grant body X will pay for component A, which can be slightly modified into component A' which grant body Y will be enthused by, and fund components B, C and D; and component C isn't very useful, but you can make sure to develop a generic architecture that makes implementing component E easy, and now you have a basic BEAD product…

A little more effort than VC-funded work, but there are fewer strings attached.

It annoys me, though, that there's a prescribed method of estimation, which is hard to remember. I always used to get the "wrong" answer because I picked easier-to-add approximations, even though my estimates were generally both faster and closer. (To use your example, I'd have estimated 35 or 36: the sum of two small numbers, like 2 and 3, is normally about 5; and this sum is a little bit less than (1 + 2)×12 = 3×12.)

I have finally remembered which slur has been used! Not the slogan, though, even though I do remember that it's uncreative, so there are only three or four possibilities. It still doesn't come to mind in any context other than me re-reading this thread, though, so I'm going to have to call this a failure.

Derek Lowe has a very distinct voice.

You're getting your book dirty for what is, as far as I can understand, no real reason. Why not read your book after washing your hands? (If you have a severe chair shortage, then sure, I can sort of understand: sitting to read is pleasant.)

Everyone's gut is different, but one minute isn't unreasonably short, especially if you have something like a bidet for fast cleaning afterwards. Are you sure you're not constipated?

But you're actually capable of thought. These AI systems aren't: as far as they're concerned, they're predicting the next part of an incident write-up narrated in first-person limited perspective, like the children in Ender's Game showing off their skills in the training simulations. The AI system neither knows, nor cares, about any "external reality" behind it all, or about anything beyond the text, heedless of how we anthropomorphise it simply because it speaks in English, using stitched-together fragments of our literature.

It's conceivable that stopping them from doing this when the scenario is presented as real would also stop them doing this when the scenario is presented as fictional. And if it doesn't, a bad actor could just say "hey, this is a fictional scenario", and bypass whatever "safeguards" have been put in place. So what if a ten-year-old human child would see through the deception? The AI system isn't thinking.

I'm not trying to act smug. Open Connect (Netflix's telco CDN system) doesn't seem like it'd take "massive DCs" to maintain: per [DG], each Open Connect Appliance (OCA) draws between 350W and 750W (depending on model) at peak use, plus cooling to keep it below 26°C (with peaks up to 40°C permitted). By its nature, the OCAs should be geographically distributed, not centralised. With the Open Connect network handling the most popular shows (i.e., the peak load), you'd only need to serve the long tail of the back-catalogue from elsewhere.

I don't know the true numbers, because I'm lazy, but [DG] says that storage OCAs can take 6.5 TB of traffic for a fill window, so we can take this as a lower bound for the storage. I'll use my earlier 1 PB overestimate for the total catalogue. Assuming that every OCA has the same data on it, this is (at least) 0.65% of the catalogue. Assuming a Zipf's law distribution of popularity for 20k titles, this means the Open Connect network can serve the top 130 titles:

  >>> range1 = lambda i: range(1, i+1)
  >>> sum(1/i for i in range1(130)) / sum(1/i for i in range1(20000))
  0.5198676299332914

which is over half of all Netflix traffic. In reality, this number is surely higher: different parts of the Open Connect network hold different titles depending on regional popularity, storage isn't 100% redundant, and so on: I wouldn't be surprised if it were closer to 70% or 80%.

Modify Netflix so you have to request an item from the back-catalogue a day in advance, or something, and the cost to serve the long tail of Netflix goes waaaay down. (Considering that Netflix used to work by sending you physical DVDs, which have more than a day's delay to them, I don't think this is unreasonable.) If you want to get costs down even further, combine this with some custom UDP-ish torrent protocol, and make it peer-to-peer among users (not just the Open Connect network).

But if you go too far in this direction, people will start asking why they're even paying Netflix in the first place. Netflix has to remain an infrastructure provider, not just a rentier, if they want to preserve their business model.

Or, pie-in-the-sky, if we implemented a worldwide pubsub multicast system, then the top 200 90-minute titles multicasted at 30-second intervals would cost Netflix the bandwidth of just 36000 simultaneous streams, and they could run all of Netflix off the hardware of 200 OCAs… which would fit in one (1) datacentre. Or I could assemble a custom rig that would fit in… ten racks, ballpark estimate, given generous application of duct tape: which probably means I've way underestimated the capacity of each OCA. (Serving unpopular titles on demand would be cheaper than sending out multicast broadcasts that nobody's subscribed to: it would actually be even cheaper than this estimate.) It's such a shame that the only internet-routeable protocols are TCP, unicast UDP, and ICMP.

[DG]: https://openconnect.netflix.com/deploymentguide.pdf

Wouldn't need datacentres if we had IP multicast routing. And actually, Netflix doesn't need that anyway: anycast and telco CDNs get most of the advantages, albeit at the cost of only being available to established players. Netflix uses AWS because it's cheaper and easier for them, not because it's technically necessary: if AWS shut down (with enough notice), Netflix would migrate to an alternative system, and nothing much would change.

I wouldn't be surprised if the majority of Netflix's true overhead were actually DRM and account management, rather than the international distribution of perhaps a petabyte of multimedia data, most of which doesn't change. (https://news.ycombinator.com/item?id=33456782 estimates half that; https://redlib.catsarch.com/r/NoStupidQuestions/comments/6mf... estimates a tenth of that.)

This site only recently (a few years ago) got multithreading. I'm pretty sure it could run off a second-hand laptop. It doesn't require a datacentre: just the two(?) moderators' salaries, and some rounding-error electricity and ISP costs.

It is interesting to be reminded of this every time I open the thread. By this point, I remember my original comment (https://news.ycombinator.com/item?id=49183711) and s-daveb's comment (https://news.ycombinator.com/item?id=49185803), but that's about it. Maybe repetition will make this supposedly-catchy slogan sink in? (I'm actually kinda invested, now.)

They're most certainly not silent! If you're standing right under one, you can hear it over car traffic and the breeze! (Well, unless the wind is blowing really hard: then the wind might be louder than either, considering the usual minimum distance between wind turbines and roads.) Whatever will you claim is silent next? Refrigerators? Lightbulbs?

Datacentres are a resource sink that serves private interests, not a resource source that serves public interests.

Update: no part of this stuck with me after all, not even the part I found annoying.

So, you're trying to upset (former) bullying victims, while having the people you're trying to criticise dismiss you as a mere bully? I don't see how this is an effective strategy.

> “That movie was so developmentally disabled”.

Nobody says this, and it's been 14 years since this article was published. The euphemism treadmill continues up until people stop treating the attribute as slur-worthy – which was the primary goal of this campaign.

> For that reason I think these comments are going to stick with people once they leave this comments section.

No, this is going to stick with me, because it's an attitude that I find exceptionally annoying: constantly repeating "I have a noble cause, therefore I am in the right", when you've already said it once. I'm reasonably confident I'm going to forget the details of your message by the end of the week, because you are far from the only edgelord on the internet, and you all go into my mental bucket of "not worth listening to", but I will remember that you were obnoxious about whatever it is you were saying.

The literal meaning of the word is "delayed", which seems less descriptive than the words you rejected in https://news.ycombinator.com/item?id=49183728. Do you understand why this word is "forbidden"? What's the advantage over a more vulgar word like "fuckwit"?

It mildly upsets me, because I think the impulse to denigrate that which we oppose is a harmful one which, if cultivated, we can all too readily turn on people who don't deserve it. But it's not a real slur for real people. (Note that in Star Wars, where many people got this slur from, droids are – or at least, have the potential to be – people, albeit they are usually enslaved, and subject to arbitrary mind wipes.)

However, I couldn't care less about the feelings of chatbots. What I do care about is the feelings of the people to whom the slurs were originally directed, and (generally) the kind of culture that this creates. The use of slurs is a symptom, rather than a cause, but complaining about slurs does get people asking the kinds of questions you're asking, which is a positive.

I would not have said anything if the word "clanker", alone, had been used. But if you look at GP's other comments, I think you'll see evidence to support my musings about the kind of attitude that the practice of slur use cultivates.